TRANSPORTATION RESEARCH RECORD 1345
Effect of Chloride and Sulfate
Contamination in Soils on
Corrosion of Steel and Concrete
OMAR SAEED BAGHABRA AL-AMOUDI, SAHEL
RASHEEDUZZAFAR, AND MOHAMMED MASLEHUDDIN
The durability performance of plain and blended cements in sab-
kha soils ( oils contaminated with h.igh concentrations of chloride
and ulfate salts) was investigated. Specimens placed in cbe sab-
kJrn soil. were obtained at periodic interval and evaluated by
cond.uct~ng trengt.h I~ s mea ureme11ts corrosfon potential
momtormg, and corrosion rate measurements at regular interval .
Results indicate that the trength reduction after 540 days of
exposure to sabkha soiJs was greater in silica fume cement mortar
specimens compared with plain and other blended cements. Re-
inforcement corrosion was, however, much lower in silica fume
cement concrete than in plain, fly ash, and blast furnace slag
cements. This indicates that use of silica fume cement in structural
component placed in abkha oils will be beneficial from the
viewpoiDt of reinforcement corro ion . Additional protective mea-
~ures . uch as application of a water-resistant epoxy-based coat-
ang, w1U have to be adopted to minimize deterioration related to
sulfate salts.
Sabkhas are saline flats underlain by sand, silt, or clay and
often encrusted with salt. These subaerial evaporative areas
border partially landlocked sea (coastal sabkhas) or cover a
number of continental depressions (inland or continental sab-
khas ). Both types form under hot, arid climates and are as-
sociated with a shallow groundwater table. Sabkha soils are
generally unconsolidated, heterogeneous, layered or unlay-
ered sedimentological framework bathed in highly concen-
trated subsurface brines. These types of soils normally have
a loose, rather porous, and permeable, gritty structure, and
their surfaces are highly hygroscopic (1). These soils typically
exist in the form of alternating cemented and uncemented
layers. The cementation is brought about by the presence of
saturated brines coupled with excessive evaporation rates (2).
Several potential geotechnical problems are associated with
sabkha oils (2). The problelllS emerge principaUy from the
high salt content in the soil and groundwater. The ulfate and
chloride concentration in the abkha soils are approximately
five to six times as high as those in typical seawater (3). The
high salt concentrations, particularly those of chlorides and
sulfates, are highly corrosive to both concrete and steel. More-
over, the conjoint prevalence of chloride and sulfate salts
0. S. B. Al-Amoudi, S. N. Abduljauwad, and Rasheeduzzafar,
Department of Civil Engineering, King Fahd University of Petro-
leum and Minerals, Dhahran 31261, Saudi Arabia. M. Maslehuddin,
Metrology, Standards and Materials Division, Research Institute,
King Fahd University of Petroleum and Minerals, P.O. Box 442,
Dhahran 31261, Saudi Arabia.
67
N. ABDULJAUWAD,
poses several unresolved questions regarding their concomi-
tant and interactive effect on the durability of reinforced con-
crete structural components placed in such an environment.
The problem is significantly magnified by the coexistence of
moisture with this high salt content, which may lead to
crystallization of salts in concrete pores, thereby leading
to its disintegration. This form of crystallization-induced
disintegration usually occurs above the water table, where
crystallization is enhanced by the prevailing high rates of
evaporation.
Geotechnical engineers usually investigate the load-bearing
and compressibility characteristics of sabkhas and are rarely
concerned with the durability performance of concrete in such
environments. This unintentional ignorance of the aggressive
service environment has led to a substantial reduction in the
useful service life of concrete structures. The normal practice
in the Arabian Gulf region is to use ASTM C150 Type V
cement to resist deterioration due to sulfate salts. The struc-
tural components placed in sabkhas, however, are subjected
to both sulfate and chloride attack. In such situations, use of
Type V cement, though providing adequate protection against
sulfate attack, fails to remove free chlorides from concrete.
To inhibit reinforcement corrosion, cement with a high tri-
calcium aluminate (C3 A) content should be used (4). Rash-
eeduzzafar et al. (5 indicated better performance by high-
~A cements in chloride environments . Although the use of
cement with high ~A phase, more than 12 percent per se, is
useful from a reinforcement corrosion standpoint, it poses the
danger of deterioration and disintegration of concrete due to
the reaction between sulfate salts and the compounds of ce-
ments. A useful approach to solve this problem seems to use
a medium C3 A cement in conjunction with suitable admix-
tures to decrease the permeability of concrete, thus retarding
the ingress of the aggressive species (6).
Mineral admixtures like natural pozzolan, fly ash, and in-
dustrial by-products like silica fume and blast furnace slag are
used in concrete to make it dense and impermeable. Whereas
there are considerable data on the sulfate resistance perfor-
mance of these cements (7-9) and some data are available
on the performance of these cements in chloride environment
(10- 13) data are fairly par e at this time on the du.rability
performance of the e cements io environment characterized
by the interaction of chloride and sulfate ions. In the absence
of data on the performance of blended cements in the chlo-
ride-sulfate ( abkha) environment a more rigorous approach
68
to the selection of appropriate composite binder components
is required before their use can be recommended in such
aggressive environments.
This investigation was carried out to evaluate the perform-
ance of three types of ASTM 150 cement (Type · I, IT, and
V) and Type I cement blended with fly a h ilica fume and
blast furnace slag in the abk ha oi l ·. The plain and blended
cement concrete and mortar specimens were made with water-
cement ratios of 0.35 and 0.50. The cement mortar specimens
were buried in sabkha soils and retrieved after 3, 6, 12, and
18 months of exposure. The performance of plain and blended
cements in the . abkha enviro nment was evaluated by deter-
mining the reduction in the compressive strength and analyz-
ing the crushed mortar specimen to measure chloride and
hydroxyl ion concentrations. The reinforcement corrosion was
evaluated by conducting accelerated laboratory studies using
simulated abkha gro undwater. T he corrosion activity was
monitored by mea ·uriog corro ion potential and corrosion
current density at regular intervals.
MATERIALS AND TECHNIQUES
Materials
ASTM C150 Type [ , Type IT, and Type V cements were used
in preparing plaiJl mortar specimens. ASTM C618 Class F fly
ash was used as 20 p rcent replacement by weight of Type I
cement to prepare portland fly ash blended cement. For the
ilica fume mortar specimen , sil ica fume was u ed as 10 per-
cent replacement by weight of Type I cement. In addition
one blast furnace lag cement containing 60 percent granu-
lated blast furnace slag, wa u. ed . A and-to-cement ratio of
2. 75 was used for mortar specimens, whereas a cement content
of 350 kg/m 3 wa used in the reinforced concrete . pecimen
used for eva luation of . teel c rro ion. The coar e aggrega t
was 19 mm maximum size crushed limestone of bulk specific
gravity 2.42 and average absorption 3.77 percent. A coarse
to fine aggregate ratio of 2.0 by weight wa kept invariant in
all the concrete mixes.
Techniques
Casting of Specimens
Mortar cube specimens 25 mm in size were used to monitor
the effect of sabkhas on compressive strength. Concrete cyl-
inders 75 mm in diameter and 150 mm in height, with a cen-
trally embedded 12 mm reinforcing bar were used to inves-
tigate reinforcement corrosion. The steel bar was coated with
an epoxy paint a t the concrete-air interface and at it end ,
which wa embedded in concrete. An effective length of 75
mm of uncoiited . tee.I was enclosed in concrete. The concrete
ingredients were mixed in a revolving mixer, whereas mortar
wa· mixed in a mortar mixer. After casting the specimens
were covered with wet burlap for 24 hr before demolding.
After demolding, the pecimen · were cured in potable water
for a further period of 13 day . Tile field specimen were
placed in plastic cages which were placed below the ground-
water table. The placement of the specimens was properly
documented , and the pit was refilled with the native oil. For
TRANSPORTATION RESEARCH RECORD 1345
investigating reinforcement corrosion, three reinforced con-
crete specimens representing similar mix constituents were
immersed in the imulated sabkha solution (15.7 percent c1 -
+ 0.5 percent so.--). Sodium chloride was used to provide
the chloJide ions and djum sulfate and magnesium sulfate
were used to provide the sulfate ions. The latter two salts
were proportioned to provide 50 percent of the sulfate con-
centration from each of them.
Reduction in Compressive Strength
Deterioration due to exposure to the aggressive environment
of the sabkha oj ls was documented by evaluating the reduc-
tion in compre ive strength. Mortar ·pecimen (25-mm cubes)
were tested in compression after 3, 6, 12 and 1 m nth using
a special 200-kN automatic machine suited for testing man
cube . At the scheduled time , the p cimens were retrieved
from the exposure site and rinsed in distilled water for about
15 min to remove any salt deposits and loose materials . They
were air dried for a period of about 24 hr and tested in
compression according to ASTM C39.
Chemical Analysis
After retrieval from the field, the specimens were cleaned of
the salt deposits, air dried, and crushed to fine powder (pass-
ing #100 sieve). The crushed portion was used to determine
t he following: (a) water soluble chlorides, by titrating against
mercuric nitrat (14), and (b) hydroxyl ion concentration, by
di · olving l 0 g of the powder sample in 500 mL di ti Iled water
and titrating the filtrate against 0.01 M sulfuric acid.
Corrosion Monitoring
Reinforcement corro ion wa monitored by measurjng cor-
rosion potentials and corro ion current density at regular in-
tervals. The corr ion potential were mea ured u ing a high-
impedance voltmeter and recording the potential with re-
spect to a saturated calomel electrode (SCE). Linear polar-
ization resistance technique was used to measure the corrosion
current den ity. T he polarization re ism.nee (Rp) was deter-
mined by conducting a linear p()Jarization ·can in the range
of ± 10 mV of the corrosion potential. A microproce sor-
ba ed potenliostat/galvano tat was used for polarizing the teel.
A stajnles steel frame placed outside the ·pecimen was u ed
as a counterelectrode, whereas an SCE was used as a refer-
ence electrode. A scan rate of 0.1 mV/sec was used. Positive
feedback technique was used to compensate for the ohmic
drop (IR) between the reference electrode and the reinforcing
bar. The corrosion current density Icacr was determined using
the Stern Geary formula (15):
Icmr = B/Rp
where
= corrosion current density (µA/cm 2 ),
Icorr
Rp = polarization resistance (ohms · cm 2 ), and
B = (~a· ~c)/[2.3(~a + ~c)].
Al-Amoudi et al. 69

13a and 13c are the anodic and cathodic Tafel constants. For w
:::J
160
COMPRESSIVE STRENGTH IN PLAIN CEMENTS
steel in concrete, values of B equal to 52 in the passive con- ...J
a:
WATER/CEl-ENT: 0. 5
~-· · ·-_._. .......a
> ,........a-..
dition and 26 in the active condition are used. Gonzalez and >-
150

r-·~
Andrade (16) have demonstrated a good correlation between a:
'?
corrosion rates determined by linear polarization resistance "- 140
technique and weight loss measurements for active and pas- LL
D
sive steel in concrete using these values of B . N

:i
130
f / "__,. . . , ~
f- _. I tJ.
l!)
z
w
0::
120 v,
f 1
RESULTS f- .r,1
en
I
w 110
> /
Strength Development u;en
w
v o TYPE V CEMENT
0::
100
The effect of the high salinity of the sabkha groundwater and CL 6 TYPE II CEMENT
:>:
D
u o TYPE I CEMENT
soil on the integrity of the cement mortar specimens was 00
evaluated by measuring the compressive strength after ex- 0 50 I 00 150 200 250 :JOO 350 400 450 500 550 600
TIME. DAYS
posure periods of 3, 6, 12, and 18 months. The compressive
strength after these exposure periods was compared with the FIGURE I Compressive strength development in plain
compressive strength before placement in the exposure site. cements (w/c: 0.50).
In normal situations, concrete strength generally increases
with the period of curing. Any reduction in strength with
increasing period of curing indicates deterioration of the in-
ternal structure, which could be caused by surface softening COMPRESSIVE STRENGTH DATA
WATER IC El-ENT: 0. 35
or expansion caused by the reaction between sulfate salts and
the cement. The compressive strength development in plain
cement mortar specimens made with a water-cement ratio of
0.5 is shown in Figure 1. The data indicate that the com-
pressive strength in mortar specimens made with Type I, Type
II, and Type V cement continues to be higher than the 14-
day value even after an exposure period of about 540 days.
The data on strength development in cement mortar speci-
mens made with Type I and Type V cements and a water-
cement ratio of0.35 are shown in Figure 2. These data indicate
an initial increase in the strength of these specimens in the
range of 120 percent over the 14-day value up to an exposure
period of about 3 months . The strength development appears
to decrease after this exposure period (i.e., 3 months). After 50 I 00 150 200 250 :JOO 350 400 450 500 550 600
an exposure period of about 540 days, the compressive strength TIME. DAYS
of these specimens was 70 to 90 percent of the 14-day value. FIGURE 2 Compressive strength development In plain
The reduction in strength in specimens made with a water cement cements (w/c: 0.35).
ratio of 0.35 may be attributed to the reaction between the
sulfate ions in the groundwater and the compounds of cement.
The strength development in cement mortar specimens made
with a water-cement ratio of 0.5 and blended with fly ash (20

!:
w
:::J
llO O TYPE I t 204 F.A. COMPRESSIVE STRENGTH DATA
percent replacement of cement), silica fume (10 percent ce- ...J tJ. TYPE I t 104 S.F. 't'/C: 0.50

""ffi' ru~~
ment replacement level), and blast furnace slag cement (60
percent granulated blast furnace slag and 40 percent Type I
cement) specimens is shown in Figure 3. These data indicate
an increase in compressive strength with period of immersion D 140
for mortar specimens made with fly ash and blast furnace slag
cement. The strength development with increasing period of I 130 P ·~..
immersion was found to be higher in fly ash cement mortar
f-
l!)
z /. ' ...... .......
specimens compared with blast furnace slag cement mortar ~ 120
f- '/ o-.... .... _ ......__ ···--··0------.........--•·-o
en
specimens. In the silica fume mortar specimens, compressive
!;!;! 110 / / A
strength reduction was indicated after an exposure period of en
I 4"'~
en l _...tJ.
about 200 days . Figure 4 shows the strength development in
~ 100
cement mortar specimens made with Type I cement and a
water-cement ratio of 0.35 and blended with fly ash and silica
a
u 00'--__._~-'-~-'-~'----L~--'--~-'---''----'-~--'--~-'-~
fume. The data indicate a trend similar to that indicated by 0 50 I 00 150 200 250 300 350 400 450 500 550 600
TIME. DRYS
the strength data in Figure 3. The strength reduction was again
observed to be higher in silica fume mortar specimens com- FIGURE 3 Compressive strength development in blended
pared with that in the fly ash mortar specimens. The strength cements (w/c: 0.50; Type I cement).
70 TRANSPORTATION RESEARCH RECORD 1345

w 160 w COMPRESSIVE STRENGTH

::J COMPRESSIVE STRENGTH ::J
-' -'
a:: WATERICHENT' 0. 50
a:: > 130
> 150
>- >-
a:: a::
D
9 '
"- 140 "-
lL lL
D D
N
N 130
0
:i :i
>--
>--
"'wz 120 "'wz
0::
0:: >--
>-- .A
tn tn
w 110 w
> >
u;tn -"' u;tn
w o TYPE V t 20Z F.R. w
0: 100 0:
TYPE I t 20Z F.A.
"'
[L
[L
TYPE V t O
t. IOZ S.F. :>:
"'D
u
D
u ~TYPE I t IOZ S.F.
90L-.~~~~~~__J"-'-~--'-~..__--''---'-~--'-~-'---'
00
0 50 100 150 200 250 300 350 400 450 500 550 600 0 50 I 00 150 200 250 300 350 400 450 500 550 600
TIME. DRYS TIME. DAYS

FIGURE 4 Compressive strength development in blended FIGURE 5 Compressive strength development in blended
cements (w/c: 0.35; Type I cement). cements (w/c: 0.5; Type V cement).

development data for specimens made with Type V cement so.lubJe chloride concentration wa 5.36, 4.62, and 5.54 per-
and blended with fly ash and silica fume and a water-cement cent in spe imens made with Type I cement blended with fly
ratio of 0.5 are shown in Figure 5. These data indicated a as h, blast furnace slag, and silica fume , respectively . The
trend similar to that shown in Figures 3 and 4. chloride conce ntration in Type V cement mortar specimen
blended with fly a b and ilica fume was 6.28 and 5.73 percent
respectively. The chlo ride conce ntration in Typel and Type
V cement made with a water-cement ratio of 0.35 wa 4.80
Chloride and Hydroxyl Ion Concentration
and 4.62 percent, respectively. The chloride concentration in
specimens made with a water-cement ratio of 0.35 and blended
The data on hydroxyl and chloride ion concentration are pre-
with fly ash and silica fume was 3.70 and 2.59 percent, re-
ented in Table 1. The OH - ion concentration in the plain
spectively. The lower chloride.concentration in specimens made
cements was in the range of 8 to 10.5 percent by weight of
with silica fume and fly ash and using a water-cement ratio
cement. The OH - ion concentration in mortar pecimens
of 0.35 is due to the densification of the paste matrix resulting
made by blending Type I cement wirb silica fume, fly ash
in the use of the mineral admixtures and the low water-cement
and blast furnace slag cement mortar was 5.54 6.56 , and 6.00
ratio. The water soluble chloride concentration in all the spec-
percent by weight of cement, respectively. The OH- ion con-
imens is, however, higher than the allowable value of 0.15
centration of mortar specimens made by blending Type V
percent suggested by ACI Committee 318.
cement with silica fume and fly ash was 4.9 and 5.3 percent,
respectively.
The water-soluble chloride content was approximately 6.28
Reinforcement Corrosion
percent by weight of cement in Types I, II, and V cements.
This indicates that the effect of cement type was not very
Figure 6 is a corrosion potential-time record for plain and
significant on water-soluble chloride concentration. The water-
blended cements made with a water-cement ratio of 0.50 and
for Type I cement with a water-cement ratio of 0.35. The
corrosion potentials were measured for three specimens rep-
resenting similar composition, and the potentials indicated in
TABLE 1 HYDROXYL AND CHLORIDE ION Figure 6 are the average of these three values. The corrosion
CONCENTRATION IN MORTAR SPECIMENS PLACED IN potentials provide a qualitative indication of reinforcement
SABKHA ENVIRONMENT
corrosion. Potential values more negative than - 270 mV SCE
OH ion Soluble Chloride c1 -1ow
Cement Type (% wt. cement) (% wt. cement) (Ratio) indicate initiation of reinforcement corrosion. The corrosion
potential - time record (Figure 6) indicates the initiation of
Type I 9.0 6.28 0.70 reinforcement corrosion in plain and blended cements in the
Type V 9.12 6.28 0.69 range of 75 to 150 days. The average corrosion potential of
Type I+ F.A. 6.56 5.36 0.82
Blast furnace s1ag 6.00 4.62 0.77 steel embedded in silica fume mortar specimens was approx-
Type I+ S.F. 5.54 5.54 1.00
Type II 10.10 6.28 0.62
imately - 270 m V, indicating no initiation of corrosion, even
Type I (w/c: 0.35) 8.74 4.80 0.55 after an exposure period of 450 days. The corrosion potentials
Type V (w/c: 0.35) J0.54 4.62 0.44
Type V + F.A. 6.46 6.28 0.97 of steel in blast furnace slag cement mortar specimens were
Type V + S.F. 4.96 5.73 l.12 more than - 270 mV at the time of immersion of these spec-
Type I + 20% F.A. 5.38 3.70 0.69
Type I + 10% S.F. 5.32 2.59 0.49 imens in the test solution. This indicates that the ASTM C876
criterion of -270 mV SCE is not useful for detecting the
Al-Amoudi et al. 71

sion viewpoint. Reinforcement corrosion was mm1mum in

1000 C TYPE I CEMENT o TYPE I t 207. F.A.
6 TYPE V CEMENT + BLAST FURNACE SL silica fume cement concrete, followed by blast furnace slag
w 900 Q TYPE II CEMENT* TYPE I t 10/. S.F . and fly ash cement concretes in decreasing order. Reinforce-
u
en • TYPE I CEMENT
> 800 ment corrosion activity was higher in plain cement concretes
f than in blended cement concrete specimens. Even an im-
700
~
_J
++ provement in concrete quality brought about by the reduction
~ 600 +
of the water-cement ratio from 0.5 to 0.35 did not significantly
>-
z
w 500 improve its performance compared with incorporation of the
>-
D
0.
mineral admixtures (fly ash) and industrial by-products (blast
400
z furnace slag and silica fume).
sen 300
D
a::
a:: 200
D
u
"' DISCUSSION OF RES ULTS
100
0 The performance of plain and blended cement in sabkba soils
0 50 I00 150 200 250 300 350 400 450 500 wa evaluated by determining trength development mea-
TIME. DAYS
uring the chloride and hydroxyl ion concentration , and eval-
FIGURE 6 Corrosion potential-time record. uating reinforcement corro ion . The compre ive trength data
indicate insigll'ificant reduction in compre ive trengtb with
period of exposure in plain cement mortar specimens made
onset of corrosion in blast furnace slag cements (17). Another with a water-cement ratio f 0.5. The effect of U1e cement
criterion normally used to detect corrosion initiation is a sub- type (i .e., C3 A variation) on strength development was not
stantial increase in corrosion activity marked by a sudden significant. The compressive strength in plain cements made
increase in the corrosion potential-time curve. Such a tran- with a water-cement ratio of 0.35 was between 75 and 90
sition has not been observed in the potential-time curve for percent of the l4-day value after an exposure period of 540
bars in blast furnace slag cement concrete. This indicates that days. Thi indicates that specimens made with a water-cement
reinforcement steel in the blast furnace slag cement concrete ratio of 0.35 may deteriorate more rapidly than those made
is in a passive condition (i.e., corrosion has not yet begun). with a water-cement ratio of 0.5. Similar results have been
The data on corrosion current density for bars embedded reported by Ben-Yair (18). In his experiments, mortar spec-
in specimens tested in this investigation are shown in Figure imens made with a water-cement ratio of 0.3 and exposed to
7. The corrosion current density was 3.0 to 3.87 µ,Alcm 2 in 3.5 percent NaCl + 0.4 percent Na2S04 solution for more
Type I, Type II, and Type V cements (w/c: 0.5). The corrosion than 8 years expanded more than specimens made with a
current density on bars in fly ash blended cement concrete water-cement ratio of 0.5. The authors attribute this behavior
was about 1.04 µ,A/cm 2 • The corrosion current density on steel to the very fine pore structure formed in the low water-cement
in blast furnace slag and silica fume cement concretes was ratio mixes compared with those made with a water-cement
0.25 and 0.08 µ,A/cm2, respectively. The corrosion current ratio of 0.50. The salts of crystallization formed as a result of
density on steel in concrete made with a water-cement ratio reaction between cement and sulfate salts are not well accom-
of 0.35 was 2.17 µ,A/cm 2 • These data indicate that blending modated in the finer pore structure in the low water-cement
of portland cement with either mineral admixtures or indus- ratio pecimen. . These salts exert considerable pressure,
trial by-products was beneficial from a reinforcement corro- resullhlg i-n greater expansion and deterioration in the low
water-cement ratio mixe than in concretes made with a high
water-cement ratio whose coarser pore structure can accom-
modate the salts of crystallization.
'E0 5.0
• TYPE I CEMENT The compressive strength data for fly ash and blast furnace
a-
(j)
4.5 • TYPE V CEMENT slag blended cements indicated an increase in strength in fly
llll TYPE I I CEMENT
'2
a::
4.0 J.87 •TYPE I t 207. F.A. ash and blast furnace slag cement mortar specimens over the
>- I'll TYPE I + 107. S.F. 14-day value, even after an exposure period of about 540 days.
>-- J. 5 ~ BLAST FLJ'lNRCE SLAG The strength development was found to be higher in fly ash
~
w
l!!I TYPE I (W/C: 0.35)
concrete than in blast furnace slag cement specimens. Mortar
D
3.0
>--
zw
specimens made with silica fume blended cements, however,
2. 5
O'. indicated a reduction in strength over the initial values after
~
u 2.0 about 200 days of exposure. The reduction in strength in silica
z fume cement mortar specimens is attributed to softening of
D 1.5
(;)
D the cement matrix due to reaction of the secondary calcium
O'. 1.0
Cl'.
D silicate hydrate ( - -H) with magnesium sulfate as ociated
u
0 .5 with the groundwaters. Thi sort of deterioration i manifested
0.0
by softening and deterioration of rhe urface skin leaving the
II F.A. S.F. BFS 0.36 aggregate expo ed to environment (scaHi1g). Such a failure
TYPE OF CEMENT
was more prominent in pecimeus made with bla t furnace
FIGURE 7 Data on corrosion current density on steel in plain slag and ilica fume cement specimen . urface caling was
and blended cements. noticeably higher in bla t furnace slag cement concret s than
72
iu plain cements in marine exp sw-e studies carried out by
one of the authors (19). Schroder and Smolzyck (20) relate
the surface deteriorati'o n to the gla s content of the slag. A
mi nimum glass content of the slag (> 61 percent) wa nec-
essary to improve the sulfate resistance of the blend.
The reduction in strength in silica fume cement mortar
specimens is attributed to the presence of Mg++ cations as-
sociated with sulfate salts. Mg++ cation is destructive to the
secondary C-S-H gel by partially converting it to a cohesion-
Iess, porous, reticulated magnesium silicate hydrate (M-S-H)
gel. The enhanced deterioration of silica fume blended cement
paste in the magnesium sulfate environment has also been
observed by Cohen and Bentur (21). They reported that bru-
cite (MH) was not traceable in silica fume blended hardened
cement pastes that had undergone deterioration due to mag-
nesium sulfate attack. MH is relatively insoluble in water and
is known to block the pores of the hardened cement paste
matrix, thereby providing a protective shield to the ingress of
sulfate ions. In view of this position, it may be hypothesized
that the absence of MH in the deteriorated silica fume blended
cements, in contrast to its presence in plain cements, presents
a more open structure, thereby making silica fume blended
cement mortar/concrete more vulnerable to magnesium-based
sulfate attack.
T he concentration of free ch loride (water-soluble chloride)
wa considerably more than a val ue of0.15 percent (by weight
of cement) recommended by ACI 318 for concrete structures
exposed to chloride-bearing waters for avoiding risk of cor-
rosion. According to Hausmann (22), the risk of corrosion is
greater when the Cl-/OH - ratio is above 0.6. The Cl - /OH-
ratio in most of the cement/mixture proportions investigated
is more than this value (see Table 1 for the values of the CI - I
OH - ratio for each of the mixes investigated). This indicates
that the changes of corrosion are elevated in reinforced con-
crete structures placed in the aggressive environment of the
sabkha. The other factor for the c rrosion to proceed is the
availability of oxygen. The rate of oxygen diffusion is signif-
icantly affected by the extent to which the concrete i san1rat-
ed with water. Since the availability of oxygen is limited in
structures placed below ground level, reinforcement corrosion
is assumed to proceed at a slower pace, even in the presence
of high concentrations of chloride ions. However, at the soil-
air interface, where abundant oxygen is available, corrosion
proceeds at a rapid race. This phenomenon appear to be of
importance in tructures submerged deep in the ground and
only a small portion of which are exposed above ground. If
these tructures are placed in highly contaminated soils, rein-
forcement corrosion i initiated at tbe soil-air interface, whereas
the buried portion remains free of corro ion because of non-
availability of oxygen. The small anode (t:he soil-air interface)
compared witb a large cathode (submerged portion) leads to
localized cono ion at the anodic sites. Reduction in the diameter
of reinforcing steel in deep foundations placed i.n sabkha oils
to the extent of more Chan 50 percent has been bserv d (23) .
The reinforcement corro ·on data indicate superior perfor-
mance by bl.ended cements in resisting reinforcement corro-
ion compared with plain cements. The corro ion rate of steel
in fly ash, blast furnace slag, and silica fume concretes was
3, 12, and 38 times lower than in plain cement concrete. The
corrosion rate of steel in blended cements made with a water-
TRANSPORTATION RESEARCH RECORD 1345
cement ratio of 0.50 was lower than in plain Type I cement
(C3 A: 8.5 percent) concrete made with a water-cement ratio
of 0.35. The corrosion current density on steel in Type I
cement (w/c: 0.35) was 2, 9, and 27 times that of corrosion
current density on steel in fly ash, blast furnace slag, and silica
fume cement concrete specimens made with a water-cement
ratio of 0.50. Furthermore, the corrosion potential-time rec-
ord does not indicate initiation of reinforcement corrosion in
silica fume cement concrete specimens, whereas reinforce-
ment corrosion was active in all the plain and fly ash cement
concrete specimens. The authors believe that the time to ini-
tiation of corrosion is an important parameter controlling the
service life modeling for structures suffering from reinforce-
ment corrosion. The other parameter used in the life predic-
tion models is the propagation period. The propagation time
remains fairly constant in all the cement types. A longer time
period for corrosion initiation observed in silica fume cement
is indicative of its usefulness in increasing the useful service
life of concrete structures compared with other cements.
The corrosion data indicate a superior performance by silica
fume blended cement concretes, implying that surface soft-
ening leading to a reduction in the compressive strength in
this cement does not accelerate the reinforcement corrosion
process. Furthermore, the magnitude of deterioration observed
in the mortar specimens is probably due to the specimen size.
It is possible that the reduction in strength due to sulfate ions
in actual structures may not be as high as that observed in
the small laboratory specimens. The surface deterioration can
be mitigated by applying a water-resistant epoxy-based coating.
CONCLUSIONS
Mortar specimen prepared using plai.11 cement (Type I, Type
II, and Type V) , and blended cements (blended with fly ash,
blast furnace slag, and silica fume) were placed in sabkha soil
for a period of more than 540 days. The performance of these
cements was evaluated by conducting compression tests, ana-
lyzing the crushed specimens for chloride and hydroxyl ion
concentrations, and monitoring reinforcement corrosion. The
conclusions from the data developed in this study are as
follows:
1. The compressive strength development in plain cements
was not affected by placement in the aggressive sabkha soils.
The compressive strength measured after about 540 days of
exposure to the sabkha environment was not affected in fly
ash and blast furnace slag cements, whereas the compressive
strength in silica fume blended cements was 70 to 80 percent
of the 14-day value.
2. The compressive trength after about 540 days of expo-
sme was 75 to 90 percent of the 14-day value in pecimens
made with a water-cement ratio of 0.35. The Compres ive
strength in cement mortar pecimens made with a water-cement
ratio of 0.5 was more than 100 percent of the 14-day value.
3. The water-soluble chloride concentration in all the cement
mortar specimens placed in the sabkha was more than the
normally accepted threshold value of 0.15 percent by weight
of cement. Similarly, the chloride/hydroxyl ion ratio was more
Al-Amoudi et al.
than 0.6 in most of the cement mixtures investigated. This
indicates that the chances of corrosion in all the cements
placed in sabkha environments are greatly increased.
4. The data on reinforcement corrosion in plain and blended
cements developed by measuring corrosion potentials and cor-
rosion current density indicate that the corrosion activity was
lower in blended cement concretes than in plain cement con-
cretes. The corrosion current density on steel in fly ash, blast
furnace slag, and silica fume concretes was 3, 12, and 38 times
lower than in plain cement concrete. The reinforcement cor-
rosion in Type I cement (C3 A: 8.5 percent) made with a water-
cement ratio of 0.35 was more than that in blended cement
concretes made with a water-cement ratio of 0.5. The best
corrosion protection was offered by plain cements blended
with 10 percent silica fume.
5. Data developed in this investigation indicate that blend-
ing of cement with silica fume protects the reinforcing steel
from chloride-induced corrosion. However, protective mea-
sures like coating the exterior surface with a water-resistant
epoxy-based coating will be necessary to protect concrete from
deterioration due to magnesium-based sulfate salts.
ACKNOWLEDGMENT
The support of King Fahd University of Petroleum and Min-
erals in the conduct of this research is gratefully acknowl-
edged.
REFERENCES
1. G . 1. E llis. Arabian alt-Bet1ri11g Soil (Sabklm) as an E11gi11eeri11g
Marerial. Report LR 523. Tran portation Road Research Lab-
ora1ory, Crowthome, Berkshire, England, 1973, 21 pp.
2. W. Akilli and J. K. orrancc. The Development and Geotech-
nical Problems or Sabkha Wilh Preliminary Experiments Oil che
Static Penetration Resistance of Cemented Sands. Quarterly Journal
of Engineering Geology, Vol. 14, 1981, pp. 59-73.
3. H. John on, M . R. Kamal , G. 0 . Pierson, and J. B. Ramsay.
Sabkhas of Eastern Saudi Arabia. Quaternary Period in Saudi
Arabia. (S . S. Al-Sayyari and J. G. Zotl, eds.). Springer-Verlag,
Austria, 1978, pp. 84-93.
4. P. K. Mehta. Effect of Cement Composition on Corrosion of
Reinforcing Steel in Concrete. Chloride Corrosion of Steel in
Concrete. ASTM STP 629. Philadelphia, 1977, pp. 12-19.
5. Rasheeduzzafar, S.S. Al-Sadoun, A. S. Al-Gahtani, and F. H .
Dakhil. Effect of Tricalcium Aluminate Content of Cement on
Corrosion of Reinforcing Steel in Concrete. Cement and Concrete
Research, Vol. 20, No. 5, 1990 pp. 723-738.
6. L. Hjorth. Cement Specifications for Concrete Exposed to Chlo-
rides and Sulfates. Proc., l111ematio11al Workshop on Durability
of Concrete Structures. Copenhagen, May 1983, pp. 229-235.
7. C. D. Lawrence . Sulfate Attack on Concrete. Magazine of Con-
crete Research, Vol. 42, No. 153, Dec. 1990, pp. 249-264.
73
8. G. Wong and T . Poole. Su/fare Resi rnnce of Morta~ Made with
Portland Cement and Blends of Por1La11d Cemeni and Pozzolan
or Slag. Technical Report SL-88-34. Department of th e Army,
Vicksburg, Mis .. 19 8 143 pp .
9. D. Stark. Long Time Study f oncrete Durability in Sulfate
Soils. George Verbeck Symposium 011 Sulfate Resistance of Con·
crete. ACJ SP-77. American Concrele Institute, Detroit, Mich.,
1982, pp. 21-40.
10. W. Hime and B. Erlin. Some Chemical and Physical Aspects of
Phenomena Associated with Chloride-Induced Corrosion. Cor-
rosion, Concrete and Chlorides . ACI SP-102. American Concrete
Institute , Detroit l987, pp. 1- 12.
11. Rashccdu1.znfar, F. H . Dakhil and K. Mukarram. J11fluence of
Cemenl Composilio11 and Co11111nt 011 rhe Corrosion Behavior of
Reinforcing Steel in oncrete. A I SP-100. American oncrete
Instit ute, Detroit 19 7, pp. 1477- 1502.
12. C. L. Page, N. R. hort, and W. R. Holden. The Influence of
Di£fcren1 Cements on hloride-Induced Corro ion of Reinforc-
ing Steel. C1m1e111 and Concrete Research , Vol. 1.6, No. I , 1986,
pp. 79-86.
13. H. K. Cook and W. J. McCoy. Influence of Chlorides in Rein-
forced Concrete. Chloride Corrosion of Steel in Concrete. ASTM
STP 629. Philadelphia, 1977, pp. 20-29.
14. Standard Methods for the Examination of Water and Waste Water .
16th edition. American Public Health Association, Washington,
1985.
15. M. S1crn and A. L. Geary. Electrochemical Polarization: I. A.
Theoretical Analysis of the Shape of Polariza1ion Corves. Journal
of £/eclrochemical Socie1y , No. 104, 1957, pp. 56-63.
16. J. A . Gonzalez ru1d C. Andrade. Quantitative Measurement of
Corro ion Rate of Reinforcing • reels Embedded in Concrete
Using Polarization Re istance Mca urcments. Werk roffe wul
Karro io11 , Vol. 29, l978, pp. 515-5 19.
17. M. Ma lchuddin , A . I. Al-Mana , H. aricimen, and M. hamim .
Corrosion of Reinforcing Steel in Concrete Containi ng Slag or
Pozzolans. ·emem 011cre1e and Aggregates, Vol. 12 No. I
1990, pp. 24-31.
18. M. Ben-Yair. The Effect of Chlorides on Concrete in Hot and
Arid Regions. Cement and Concrete Research, Vol. 4, No. 3,
1974, pp. 405-416.
19. A . R. AJ-Rabiah, Ra heeduzzafar, and R. Baggot. Influence of
Cc.:menl Type and Mix Composition on oncre1e Deterioration
in 1he Marine Arabian Gulf Environment. Proc., Third l11ter-
1wrio11al Conference 011 Deterioration and l~epnir of Rei11forcetl
Concrete i11 the Arabian G11/f. Oc1. 1989, Bahrain , pp . 493- 528.
20. F. Schroder and H. G . Smolczyk . Blasl Furnace lags and Slag
Cements. Behavior in Aggressive olutions. Proc. 5rh /111erna-
tio11al Symposium 011 rite Chemistry of Cement, Vol. 4, Tokyo,
1968, pp. l J-199.
21. M. D. Cohen und A. Bentur. Durability of Porlland Ccment -
Silica Fume Pastes in Mngncsiurn Sulfate and . odium ulfate
olutions. AC/ MarerialsJoumal, May- June, 1988, pp. 148- 157.
22 . D. A. Hausman. teel Corrosion in Concrete , How Does It
Occur? Materials Protection Vol. 6, No. 11 , 1967, pp. 19-23.
23. M. Maslehuddin, H. aricimen. A. I. Al-Mana , and M. hamin.
Performance of Concrete in a High Chloride-Su/fate Environment .
ACI SP-122. American Concrete Institute, Detroit, Mich. , 1990,
pp. 419-494.
Publication of Lhis paper sponsored by Committee on Soil and Rock
Properties.